Johannesburg, 23 Jan 2006
The biggest challenge when deploying intelligence systems is to create a common understanding between the IT team and the business. It is also critical that teams themselves share mental models and a common language.
This is the view of Johann Heymann, CEO of Fifth Discipline Group, who addressed delegates at the recent SAS Forum, SA`s premier BI user group event.
"Mental models are one`s way of looking at the world, and establish how one thinks," said Heymann. "Everyone forms mental models of everything and these may be very counter-productive in an IT engagement."
He explained that models are, by definition, a simplification of the truth, and that mental models are more unstable and difficult to work with than mathematical models.
"People form perceptions and stereotypes which can be dangerous to an IT implementation, and cause continuous misunderstanding between individuals on assignments," he said.
Mental models come into play as people obtain data, add meaning, make assumptions, draw conclusions and ultimately start forming beliefs. When deploying intelligence systems it is thus critical that teams share mental models.
"This does, however, take time and effort," said Heymann. "Teams must actively grapple with questions like `how do we understand this problem?` and `how do we go forward?`
"A shared vision is critical, but impossible if the team has neither the propensity to learn, nor similar mental models."
In any engagement it is critical to agree on a common language at the outset. In its first phase, the SAS intelligence platform methodology, for example, proposes a glossary of business and IT terminology. Per engagement, one must develop and agree upon terminology. It is even more critical to agree on definitions when technology is new or in multi-vendor environments.
Obstacles to common understanding include insufficient direct interaction between the IT team and the business, lack of sufficient commitment from leadership, no common idiom, misinterpreted metaphors and little commitment to continuous learning.
Research has found that approximately 70% of the success criteria in successful BI implementations involve a significant measure of understanding and involvement by the client. "Mutual understanding is absolutely key in successful project delivery," said Heymann.
In the IT realm projects tend to be iterative because IT by nature is a development process. Thus the typical project management focus on on-time, within budget delivery is misguided.
"Ultimately, projects are done to achieve benefits to the organisation, which has a lot more in mind than merely reaching the deliverables," said Heymann.
In any IT project, it is given that one will have to re-scope often. The client learns, and as he does so, the scope changes. Ultimately, one wants to create value for clients.
IT implementations are also fraught with `wicked` problems - those for which there is no definite statement, and which present interlocking challenges. For example, stakeholders have different views of how value is created, as well as hidden agendas, while IT teams have limited resources.
To manage wicked problems, one has to think creatively, use a systems thinking approach, and have a common language.
Systems thinking has to do with cause and effect. "The reality is that there are curved balls all the time during implementation," said Heymann. "Systems thinking addresses these changing inputs."
The scrum technique is a useful one. Priorities are set on a regular basis, for example monthly, and the team implements on these priorities. At the next meeting, a new priority list is built so that as things change, items are re-prioritised.
"Wicked problems are common in our daily and business lives as well as in the IT environment," said Heymann. "Often one cannot fully understand these problems, so one has to build on those issues one can control and understand."
Share